A linear model with two covariates seen at the same time in terms of both (3d).
US Real interest ex-post trimester time series from 1961 until 1986, with empiric M-fluctuation process based on the scores of a fitted model (break/change points).
The curve fits obtained through each gradient descent iteration until convergence, in red.
Sequential Bayesian learning of a simple linear model with one covariate.
Clustering results of the DbScan algorithm for different minpts parameter values.
Clustering results of the DbScan algorithm for different eps parameter values.
A: Deviance contour of the quadritic approximation for a bivariate Gaussian with zero mean;
B: Quadratic approx. added into the original likelihood-based deviance contour.
Deviance contours for a bivariate Gaussian with zero mean.
Deviance contours with the parameters estimate at each optimization algorithm function evaluation.
Posterior mean and 95% credible interval of a random walk model of 1st order for smoothing splines without intercept.
